[Linux-HA] 2.1.2 and failover of colocated resources

Andrew Beekhof beekhof at gmail.com
Wed Sep 19 03:04:37 MDT 2007


On 9/19/07, Klemens Kittan <kittan at cs.uni-potsdam.de> wrote:
> Am Tuesday, 18. September 2007 17:16 schrieb Andrew Beekhof:
> > On 9/18/07, Klemens Kittan <kittan at cs.uni-potsdam.de> wrote:
> > > > shortly I'll be releasing a revised implementation (including
> > > > documentation!) of colocation which will make it much more intuitive
> > > > and remove the need for hacks like symmetrical=true
> > > >
> > > > if anyone wants to try it sooner rather than later, grab the latest
> > > > from http://hg.linux-ha.org/dev and ping me for the current version of
> > > > the docs.
> > >
> > > Thank you for the link, we checked out the latest version (2.1.3).
> > > Compiling and starting worked well.
> > >
> > > We did 2 types of starting the heartbeat:
> > >
> > > a) without symmetrical="true"
> >
> > you mean on the colocation constraint right?
> >
>
> Yes.
> <rsc_colocation id="group_on_drbd" from="group" to="ms_drbd" to_role="Master"
> symmetrical="true" score="INFINITY"/>
>
> > > The two systems started as master and slave and the group is started at
> > > the master node. But still when one service in this group fails, the
> > > complete group is deactivated and the state of master and slaves leaves
> > > unchanged. Normally the slave should become master and the group should
> > > start on this node.
> >
> > can you attach the complete CIB when the cluster is in this state?
> >
>
> Attached cib.xml is the output of cibadmin -Q in the situation described
> above. Hope you can see something in there...

ah yes, this is the reason i've not released it yet... i'd not
verified the master/slave interactions yet

the placement of the masters needs to include the same sort of
logic... shouldn't be hard, i'll try and do that today



More information about the Linux-HA mailing list